Generative AI Engineer · Agent Systems · Data Science
Bengaluru, India · IST (UTC+5:30) · Open to remote opportunities
Portfolio / LinkedIn / Kaggle / Email me
I build LLM applications, multi-agent workflows and data products with Python, TypeScript and cloud services. My work spans retrieval-augmented generation, model evaluation, voice interfaces and the software around them: APIs, state, observability and usable frontends.
I'm interested in remote Generative AI Engineer, Applied AI Engineer and AI-focused Python Developer roles, as well as data science opportunities where models become useful products.
- Remote experience: Freelance AI Agent Specialist at Mindrift, working on RLHF evaluations and Python code-generation assessment.
- Education: Pursuing the online Master's in Data Science at the University of Pittsburgh, expected February 2027. B.Tech in Computer Science & Engineering, Techno International New Town, 2024.
- Hackathons: My team’s VyaparGyan project was a Top 36 finalist at AI for Bharat, powered by AWS with Hack2skill. Background and achievements →
- Experiments: Published a Kaggle writeup on fine-tuning Gemma 3 for verifiable reasoning.
A selection of my independent projects, capstone work and team hackathon builds.
| Project | What I built / contributed to | Engineering focus |
|---|---|---|
| Cred Domain Support Agent | A lending-support capstone that retrieves cited policy answers, checks synthetic loan records and routes escalation recommendations. | Python · LangGraph · Chroma · FastAPI — input/output guards, checkpointing, evaluation and an offline mock mode. |
| VyaparGyan | A team retail platform combining conversational commerce, multilingual voice workflows and AI-assisted inventory operations. | TypeScript · Next.js · AWS Lambda · DynamoDB · Bedrock — event-driven workflows, role-based access and infrastructure as code. |
| Omni-Lab | A multimodal learning workspace with document-to-slide generation, live voice tutoring, flashcards and quizzes. | React · TypeScript · Gemini — voice interruptions, structured model output and synchronized UI state. Project story → |
| Bruhworking / NexusFlow | A document workflow with planner, researcher, writer and reviewer agents. | Python · FastAPI · Google ADK · Cloud Run — SSE progress streaming, revision loops and per-agent usage tracking. |
| Netflix Content Analytics | A reproducible catalog-analysis pipeline from ingestion and cleaning to exploration, modeling and a dashboard. | Python · PostgreSQL · Pandas · Streamlit — data preparation, statistical analysis and interactive communication. |
More data science: Employee attrition · Stock forecasting · Drug side-effect analysis
Public pull requests in other people’s repositories. Merged cards show contributions merged by someone other than me; open cards show upcoming work awaiting a decision.
Explore all my public contributions · Cards refresh automatically every 6 hours.
| Area | Tools and methods |
|---|---|
| AI & agents | LangGraph, LangChain, Google ADK, Gemini, RAG, Chroma, LLM evaluation |
| Data & ML | Python, SQL, Pandas, scikit-learn, exploratory analysis, feature engineering |
| Applications | FastAPI, TypeScript, React, Next.js, Streamlit |
| Cloud & delivery | AWS, Google Cloud, Cloud Run, Docker, Git, GitHub Actions |
Google AI Professional Certificate · Python for Data Science and AI · Data Science Methodology · Tools for Data Science V2
Issued by Coursera; the three data science badges are authorized by IBM. View my public Credly wallet →
Selected Claude Academy course-completion badges. Each card opens its official verification page.
All 20 Anthropic course-completion links
- AI Fluency for pK–12 educators
- AI Fluency for nonprofits
- AI Fluency for small businesses
- AI Fluency for students
- Building effective human-agent teams (beta)
- Building with the Claude API
- AI Fluency: Framework & Foundations
- AI Fluency for educators
- AI Capabilities and Limitations
- AI Fluency for Creative Work
- AI Fluency for Builders
- Claude with Amazon Bedrock
- Claude with Google Cloud's Vertex AI
- Claude Code in Action
- Deploying Claude Enterprise with Confidence: The five decisions that shape your rollout
- Claude Platform 101
- Introduction to Model Context Protocol
- Claude Code 101
- Introduction to Claude Cowork
- Claude 101
Highlights from the 20 badges on my Kaggle profile, checked October 9, 2026.
 5-Day AI Agents Intensive Course with Google |  Research Competitor |  Code Submitter |  Competitor |
 Python Coder |  Dataset Creator |  API Notebook Creator |  API Dataset Creator |
See all 20 earned Kaggle badges
- 5-Day AI Agents Intensive Course with Google
- Research Competitor
- Code Submitter
- Competitor
- Code Tagger
- Vampire
- API Dataset Creator
- API Notebook Creator
- Colab Coder
- Code Forker
- 1 Year on Kaggle
- Dataset Creator
- Competition Modeler
- March Mania Competitor
- Notebook Modeler
- Code Uploader
- Python Coder
- Kaggle Community Member
- 7 Day Login Streak
- Agent of Discord
On Kaggle, I explore model training and applied ML. My Gemma 3 reasoning project investigates fine-tuning with Tunix / JAX and GRPO, with a public writeup and project notebooks. Read the experiment →
On my portfolio, you can find project context, experience, education and credential links.
A snake making its way through my GitHub contributions. Refreshed daily.
View contribution activity · Powered by snk
Hiring for a remote AI or data role? I'd love to discuss your product, the engineering challenges and where I can contribute.
Email basuoikantik@gmail.com → · Connect on LinkedIn · Explore my portfolio